June, 1955
EFFECT OF C, N, 0 AND S ON
IRON SURFACE TENSION
LIQUID
557
SURFACE TENSION AT ELEVATED TEMPERATURES. 11. EFFECT OF Cy N, 0 AND S ON LIQUID IRON SURFACE TENSION AND INTERFACIAL ENERGY WITH A1203 BY F. A. HALDEN AND W. D. KINGERY Ceramics Division,' Department of Metallurgy, A1 assachuselts Institute of Technology, Cambridge, Massachusells Received J a n u a r y 16,1066
The eRect of additions f: C, N, 0 and S on the surface tension of liquid iron and its interfacial energy with A1203 havc been determined a t 1570 . Surface tension of pure iron was found to be 1720 dynes em.-'. Oxygen and sulfur form monolayers on the surface a t concentrations below 0.1%. Surface activity decreases in the order S > 0 > N > C.
Introduction Previously it was qualitatively observed that small amounts of oxygen markedly lowered the surface tension of liquid metals.2 Similarly, sulfur has been reported as surface-active in liquid copper,3.and silicon has been found to be selectively adsorbed at a liquid iron-aluminum oxide interfacea4 In general, reported data for liquid metal surface tension a t elevated temperatures have concerned systems in which minor impurities were not eliminated. There have been no previous investigations df the quantitative effects of 0, N and S in liquid iron. Reported data for C are not in agreement. Data for pure iron vary from S5802 t o 13805 dyne/cm. Data for about 4-5% C vary from 17206 and SSO07 to GOO8 dyne/cm. In the present investigation, the effects of these materials on surface and interface energy for liquid iron in contact with have been determined.
Experimental
ensure a uniform advancing contact angle), weighed, rinsed with acetone and placed with glass tongs on an A1203 plaque prepared by calcining aluminum hydroxideo (J. T. Baker, reagent grade), pressing and firing a t 1850 . Plaque surfaces were polished with a fine diamond lap, washed, dried and used without touching the surfaces, in order to avoid any ossible contamination. After placing in the furnace and Eveling, the systemwas heated to 1000" under vacuum (0.005 p ) and then heated to 1570' for measurements in.0.5 atm. pf helium. Hospital grade helium wafi purified with a liquld nitrogen cold trap, CuzO a t 400" to oxidize any reduced gases, Mg chips a t 600" and sponge Ti a t 1000° t o remove oxidized gases, and activated charcoal a t -200". Iron-nitrogen compositions were studied using a flow system of purified nitrogen and argon mixed in a flow-meter with a total flow rate giving a displacement of one foot per minute in the furnace. Measurements were made with both increasing and decreasing nitrogen partial pressure in order to ensure that equilibrium was obtained. Nitrogen activity was determined from the partial pressure and Sievert's law
a = 0.03934Px (1) In view of the effect of small amounts of oxygen and sulfur, alloys were analyzed for 0, S, N and C before and after surface tension measurements. Compositions and experimental measurements are shown in Table I. No change was found before and after measurements except in the one case indicated.
The sessile drop method previously described2 has been employed to measure both surface tension and contact angle. The calculated uncertainty of results due to unTABLE I certainty of measurements is &2%. Experimentally, it is found that maximum deviations obtained vary between RESLL,TH I'UU SURFACE TENSION ANL) 1 and 3% for a given composition. ~IEASUREMENTS High-purity, vacuum-melted iron (Vacuuin Metals Corporation Ferrovac E Ingot) was employed as the base Surface inetal for all compositions prepared. This material has the tension, Couiposition, following impurities: 0.0031% C; 0.0072 0; 0.00051 K; dyy Oxygen Carbon % Sulfur c m .0.005S; < 0.003 Al; < 0.006 Co; < 0.001 Cu, R4n; < 0.01 Ni, Pb; 0.01 Si; < 0.0005Sn; < 0.003 Mo. Iron alloys 0 ,OOOG 0.027 0.005 1717 with C, 0 and S were prepared by melting in purified helium .OOi7 0,009 0,005 1632 after outgassing a t red heat in uacuo. The alloying agents .020 ... 0,005 I541 were placed in the center of iron ingots machined to the .041 ... 0,009 1362 shape of the alumina crucible in order to prevent reaction with the container. Alloying additions were spectroscopic .07a ... 0,010 1151 grade carbon (National Carbon Corporation), ferric oxide . 0006 0.027 . . . 1717 (Baker & Adamson reagent grade), and ferrous sulfide ,0007 0.47 ... 1701 prepared in this Laboratory by passing sulfur vapor (Mallinc1708 .0018 ... 3.39 krodt, ppt. sulfur) over pure iron turnings in a purified atmosphere. From the resulting alloy ingots, samples .0205 0.003 1281 0,085 (2 9.) were prepared in approximately spherical shape (to 0.36 .0215 ,004 970 2 .oo ,0398 ,004 707 ( 1 ) With funds from the U. 8. Atomic Energy Commission under Contract Number A T (30-1)- 1192. .0079 ,004 1.70 70;
(2) W. D. Kingery and M. Hurnenik, Jr., THISJOCRNAL,67, 369 (1953). (3) C. F. Baes and €1. H. Kellogg, J . Metals, May, 643 (1953). (4) W.D. Kingery, J . Am. Cer. SCC., 37, 42 (1954). (5) F. Becker, F. Harders and H. Kornfeld, Arch. Eisenhllttenwesen, 20, 363 (1949). (0) J. Kevarian, Thenis, Metalluray Department, hf. I . T., 1954. (7) P. I